Chris Dixon pfp
Chris Dixon
@cdixon.eth
“Tastemates”: somewhere out there is someone who likes all the same [books/movies/music etc] as you. Have an algorithm find that person and surface the things they love that you haven’t discovered yet. Makes the recommendations more trustworthy.
24 replies
0 recast
0 reaction
Ben O’Rourke pfp
Ben O’Rourke
@bpo
Oooh great idea
0 reply
0 recast
0 reaction
Alessandro pfp
Alessandro
@azeni
@perl
0 reply
0 recast
0 reaction
Alex Reinstein pfp
Alex Reinstein
@asr
@perl
0 reply
0 recast
0 reaction
haltakov.eth pfp
haltakov.eth
@vlad
One of the problems for these general "tastemates" is having a dataset across all these modalities: books, movies, music. Maybe achievable in the future if we start using free, soulbound tokens for all kind of these activities and build our "taste identity" on the blockchain
1 reply
0 recast
0 reaction
haltakov.eth pfp
haltakov.eth
@vlad
I want to have something like this for restaurant/sights recommendations when travelling. In theory, you should be able to build a graph like A and B both liked restaurant X in Paris, A liked restaurant Y in Berlin, recommend Y to B when visiting Berlin.
0 reply
0 recast
0 reaction
Kostas Christidis pfp
Kostas Christidis
@kostas
criticker.com works exactly on that principle and allows you to view your tastemates' profiles too
0 reply
0 recast
0 reaction
Ben Scharfstein pfp
Ben Scharfstein
@scharf
This is pretty much what google discover (content recommendations in the google app) does generalized to more than one tastemate. Started with articles but also longer form content, podcasts, videos etc. It’s also what Netflix, good reads, Spotify all do on a more local basis.
0 reply
0 recast
0 reaction
0x pfp
0x
@aizimuthal
Is that just a Hunch or… 😉
0 reply
0 recast
0 reaction
Dave Morin pfp
Dave Morin
@davemorin
This was the original genius of last.fm’s musical neighbors. One of the first social use cases I fell in love with.
2 replies
0 recast
0 reaction
Dan | Icebreaker pfp
Dan | Icebreaker
@web3pm
In practice: Hey, looks like you and I share 99.9% common interests. What’s wrong with you? Why don’t you like XXX?! Thanks to evolution we are incredibly adept at finding ways to differentiate ourselves, especially from those most like us
0 reply
0 recast
0 reaction
Jun Gong pfp
Jun Gong
@jun
@perl
0 reply
0 recast
0 reaction
sree 🎩 pfp
sree 🎩
@sree
Or the exact opposite would work too, especially in the case of finding a partner. Someone who dislikes the same things as you.
0 reply
0 recast
0 reaction
gabe pfp
gabe
@gabe
I want the inverse: to find someone as dissimilar to me as possible
0 reply
0 recast
0 reaction
capers pfp
capers
@cap
I like the search version of a related idea, where results are ranked or weighted by proximity to people you follow
0 reply
0 recast
0 reaction
Peter Kim pfp
Peter Kim
@peter
@perl
0 reply
0 recast
0 reaction
zico pfp
zico
@zico
would love something like this. puts context around "people who bought x also bought y." TaaS
0 reply
0 recast
0 reaction
Landon pfp
Landon
@lndnnft
@perl
0 reply
0 recast
0 reaction
chriscantino pfp
chriscantino
@chriscantino
If we can generate recommendations and social feeds based on faves plus wallet activity (as much as users or anons are willing to opt in), that’s a start? Tie privacy prefs and affinities in at the ENS level?
0 reply
0 recast
0 reaction
Alessandro pfp
Alessandro
@azeni
@perl
0 reply
0 recast
0 reaction